- From: John Bradley <john.bradley@wingaa.com>
- Date: Wed, 6 Aug 2008 14:43:35 -0700
- To: noah_mendelsohn@us.ibm.com
- Cc: "Booth, David (HP Software - Boston)" <dbooth@hp.com>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, Julian Reschke <julian.reschke@gmx.de>, www-tag@w3.org
- Message-Id: <BDF73301-681F-4243-9B08-63DE753FB8A6@wingaa.com>
Noah, On 6-Aug-08, at 10:28 AM, noah_mendelsohn@us.ibm.com wrote: > We agree, or at least we appear to: the empty string is legal as an > IRI > Reference and as a URI Reference. The empty string is also a member > of > the value and lexical spaces of the XSD anyURI datatype. The the > empty > string is also legal as the value of a namespace declaration in XML, > that > string is a distinguished value that is in fact used to cancel the > corresponding prefix binding. Accordingly, there are no namespaces > named > with the empty string. > >> Relative and "empty string" IRI are certainly valid IRI, the >> question is if they are valid "anyURI" in XML 1.1? > > I assume you mean XSD 1.1? The specification says [1]: Yes XSD 1.1 Draft 20 > > > "The ·lexical space· of anyURI is the set of possibly empty finite- > length > character sequences." Earlier it says: > > "[Definition:] anyURI represents an Internationalized Resource > Identifier Reference (IRI). An anyURI value can be absolute or > relative, > and may have an optional fragment identifier (i.e., it may be an IRI > Reference). This type should be used when the value fulfills the > role of > an IRI, as defined in [RFC 3987] or its successor(s) in the IETF > Standards > Track. " > > So, >any< sequence of characters, including the empty sequence, is a > legal > xsd:anyURI. The specification says that anyURI >should< be applied > when > the string represents an IRI reference. Conformance to IRI ref. > syntax is > thus strongly encouraged, but not enforced by the datatype. > >> I discounted that as perhaps being non-normative given the >> draft status and the other document. > So I conclude that contrary to http://www.w3.org/TR/xml-names11/ #reluri , the recommendation http://www.w3.org/2000/09/xppa is not making it into XSD 1.1? Perhaps some mention should be made in http://www.w3.org/TR/xml- names11 that it is no longer authoritative. I guess that makes the answer to Julian's original question. No http://www.w3.org/TR/xml-names11/ is not "the namespaces specification for XML 1.1", http://www.w3.org/TR/xmlschema11-2/ should be used. Regards John Bradley OASIS IDTRUST-SC http://xri.net/=jbradley 五里霧中 > If we're discussing XSD 1.1, then [1] is as authoritative a source as > you'll find. The only Recommendation-level version of XSD is 1.0 [2]. > That version did try to enforce conformance to (pre-) IRI syntax, but > clearly allows for relative and thus for empty forms. I think there's > very little question that XSD anyURI, whether 1.0 or 1.1, allows > for IRIs > (perhaps modulo some edge cases in 1.0 as you say that you and DaveO > uncovered some), allows for relative IRIs, and thus allows for the > empty > string. We agree that the empty string is not and never has been > usable > as the name for an XML namespace. > > Noah > > [1] http://www.w3.org/TR/xmlschema11-2/#anyURI > [2] http://www.w3.org/TR/2004/PER-xmlschema-2-20040318/#anyURI > > > -------------------------------------- > Noah Mendelsohn > IBM Corporation > One Rogers Street > Cambridge, MA 02142 > 1-617-693-4036 > -------------------------------------- > > > > > > > > > John Bradley <john.bradley@wingaa.com> > 08/06/2008 12:32 PM > > To: noah_mendelsohn@us.ibm.com > cc: "Booth, David (HP Software - Boston)" <dbooth@hp.com>, > "Henry S. Thompson" <ht@inf.ed.ac.uk>, Julian Reschke > <julian.reschke@gmx.de>, www-tag@w3.org > Subject: Re: Boeing XRI Use Cases > > > Hi Noah, > > I don't claim to e an XML 1.1 expert. > > I was looking at Sec 2.2 of xml-names11: > http://www.w3.org/TR/xml-names11/ > >> 2.2 Use of IRIs as Namespace Names >> The empty string, though it is a legal IRI reference, cannot be used >> as a namespace name. >> >> The use of relative IRI references, including same-document >> references, in namespace declarations is deprecated. >> >> Note: >> >> This deprecation of relative URI references was decided on by a W3C >> XML Plenary Ballot [Relative URI deprecation]. It also declares that >> "later specifications such as DOM, XPath, etc. will define no >> interpretation for them". >> > > I should have used "empty string" rather than null. > > There seems to be a conflict between the two documents. > > The reference to relative IRI XML schema re 3.3.18 of xmlschema11-2 is > in a [Definition:] > > I discounted that as perhaps being non-normative given the draft > status and the other document. > > Relative and "empty string" IRI are certainly valid IRI, the question > is if they are valid "anyURI" in XML 1.1? > > It is worth someone having a look at, but probably someone closer to > it than me. > > However I don't think it is particularly relevant to XRI though I > suppose that relative XRI would also be prohibited according to xml- > names1.1. I don't think I am going to lose sleep over that:) > > The XRI specs don't rely on any new functionality in XML 1.1. > > Regards > John Bradley > OASIS IDTRUST-SC > http://xri.net/=jbradley > 五里霧中 > > > > > On 6-Aug-08, at 7:29 AM, noah_mendelsohn@us.ibm.com wrote: > >> John Bradley writes: >> >>> The definition of "anyURI" changes slightly to include the >>> mapping of http: IRI according to RFC3987 This allows for the >>> ireg-name component to be mapped via RFC3490 for schemes using >>> domain names. >> >> OK, good. >> >>> In XML schema 1.1 that is almost done [http://www.w3. >>> org/TR/xmlschema11-2/] >> >>> [...] allows almost all http: scheme IRI to work as "anyURI", >>> relative and null IRI are excluded so it is still a sub set, >>> though a much larger one than before. >> >> Hmm, I don't see where relative or "null" are excluded. The >> specification for that datatype says [1]: >> >> "[Definition:] anyURI represents an Internationalized Resource >> Identifier Reference (IRI). An anyURI value can be absolute >>or >> relative<<, and may have an optional fragment identifier (i.e., it >> may be >> an IRI Reference)." >> >> Was there something else you meant when you said that "relative [is] >> excluded"? I read this as explicitly allowing relative. As to >> "null" >> IRIs, the word null does not appear in RFC 3987 [1] or in RFC 3986 >> [2] for >> that matter, but RFC 3987 includes the following grammar: >> >> IRI-reference = IRI / irelative-ref >> >> irelative-ref = irelative-part [ "?" iquery ] [ "#" ifragment ] >> >> irelative-part = "//" iauthority ipath-abempty >> / ipath-absolute >> / ipath-noscheme >> / ipath-empty >> >> ipath-empty = 0<ipchar> >> >> This seems to me to indicate that an IRI Reference, as required by >> XML >> Schema 1.1, can indeed be an ipath-empty, I.e. the null string. Am I >> still missing something? >> >> Noah >> >> [1] http://www.ietf.org/rfc/rfc3987.txt >> [2] http://www.ietf.org/rfc/rfc3986.txt >> >> -------------------------------------- >> Noah Mendelsohn >> IBM Corporation >> One Rogers Street >> Cambridge, MA 02142 >> 1-617-693-4036 >> -------------------------------------- >> >> >> >> > > >
Attachments
- application/pkcs7-signature attachment: smime.p7s
Received on Wednesday, 6 August 2008 21:44:23 UTC